Machine Learning
Avoid these errors that are easily missed from machine performance flow – Part 2 | By Thomas a Dorfer | Jan, 2025
Using data unavailable during predicting and mixing magical numbers and real numbers
Welcome to another program in this series of easily missed the flow of a machine reading work! For those who have not yet learned the first, this is part of the most focused series of the potential process but have the highest power of the functional degeneration when it ends into our development pipe. .
In the first article, we examine common snares such as misuse of numeric identifiers, mistreating data separation, and filling the model in rare features.
In this type, we will continue to explore certain errors related to data management, focus on the following two articles:
- Data training is not available during prediction
- Matching magic numbers and real numbers
